We back the builders securing what AI makes possible.

Brightmind Partners announces their investment in Artemis, a new security operations platform founded by Shachar Hirshberg and Dan Shiebler (former leaders at AWS and Abnormal Security) that aims to reimagine SIEM for the AI era. The company introduces "adaptive protection," an approach that automates detection, investigation, and response by embedding deep enterprise context into an autonomous detection lifecycle, eliminating false positives and evaluating threats as complete attack stories. Unlike traditional SIEMs built on centralized data ingest and retention economics, Artemis offers architectural flexibility that works across multiple SIEMs, datastores, or its own storage, avoiding vendor lock-in. Brightmind praises the team's rapid execution—building in six months what the SIEM space has attempted for over a decade—along with their strong customer-obsessed culture.

Stephen argues that while Mythos (an AI security model) giving 50 companies a 90-day head start to find vulnerabilities sounds promising, it's largely illusory — enterprises are already drowning in unpatched critical vulnerabilities and the remediation pipeline is far too slow and contentious to meaningfully close the gap in time. On day 91, when Mythos is broadly released, whatever defensive advantage was gained evaporates and attackers gain the same powerful exploit-generation capability. He recommends CISOs restructure their security orgs entirely — retiring traditional scanning teams in favor of one massive remediation-focused team, letting AI handle identification while humans focus solely on fixing. His deeper fear is that these capabilities will trickle into small, offline, guardrail-free local models, democratizing exploit generation in a way that creates systemic risk far beyond what any centralized tool with access controls can contain.

Brightmind invested in Onit Security, founded by Ofer Amitai, Elad Ben Meir, and Tom Winter, with Amitai's firsthand experience watching a vulnerability backlog lead to the compromise of his prior company Portnox serving as the driving force behind an entirely new approach to exposure management. Traditional vulnerability management has been fragmented across siloed scanners and homegrown workarounds, consistently failing enterprises due to limited asset context, inconsistent prioritization, and the inability to act without unintended consequences. Onit's decision-based, AI-native platform unifies business context with risk signals across the enterprise, embedding operator feedback directly into the engine to ensure autonomous, context-aware remediation that closes gaps at machine speed. With adversaries achieving breakout times as short as 27 seconds using AI, Brightmind sees Onit as the foundational layer for autonomous vulnerability management while also delivering meaningful productivity gains for IT and engineering teams.

Brightmind invested in Vijil, founded by Vin Sharma, Zdravko Pantic, Radina Mihaleva, and Tim Rudner, a team combining deep AWS infrastructure pedigree, enterprise GTM experience, and world-class AI research credentials from institutions including Oxford, Yale, and Caltech. Enterprises are actively deploying AI agents but lack the infrastructure to verify they are safe, reliable, and production-ready — a gap that content filters, static guardrails, and first-generation AI security tools have consistently failed to close. Vijil addresses this with a unified trust platform built around three compounding modules: Diamond for continuous, bespoke agent testing; Dome for policy-driven runtime guardrails with real-time enforcement; and Darwin, which uses multi-agent reinforcement learning over production telemetry to continuously harden agent behavior. With early enterprise customers including DigitalOcean, SmartRecruiters, and DuploCloud, partnerships with AWS, Google, and Groq, and recognition as a Gartner Cool Vendor for AI Security, Brightmind believes Vijil is positioned to become the defining trust infrastructure of the agentic era.

Brightmind invested in SGNL's Series A in January 2025, backing a platform purpose-built to close the long-standing gap between security operations and access management by embedding real-time context into every access decision. SGNL's dynamic, continuous identity architecture stood in contrast to the static approaches of legacy PAM vendors, positioning the company to meet the access governance demands of the agentic AI era. CrowdStrike has now agreed to acquire SGNL, integrating its access management capabilities with CrowdStrike's industry-leading identity threat detection to deliver a consolidated platform the firms are calling "threat-driven continuous identity." The combination gives CrowdStrike a differentiated end-to-end identity stack and marks a strong outcome for Brightmind's early conviction in the SGNL team.

Brightmind invested in SGNL based on a founding team with half a century of identity security experience, having previously built and sold companies to Google and pioneered the invention of CAEP. The firm identified a fundamental architectural gap in the incumbent PAM market, where static, pre-configured permissions leave enterprises exposed to insider threats and credential abuse that legacy vendors lack the architecture to address. SGNL's real-time, context-aware approach eliminates standing privileges by acting as a central nervous system for identity — ingesting risk signals from across the enterprise stack and enforcing dynamic, policy-driven access at adversary speed. With large enterprises actively pursuing identity transformation projects to replace 20-year-old technologies, Brightmind saw SGNL as well-positioned to become the gold standard for privileged identity management.